fabi/Snappy
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
423 Commits 2 Branches 0 Tags
aa5f5f865075b717c1532559ee819da57280fee0
Commit Graph

423 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Danila Kutenin
aa5f5f8650 Internal change 
PiperOrigin-RevId: 626960053
 2026-05-09 18:52:47 +00:00
Connal de Souza
2ac1356d4d Optimize ExtractOffset() by avoiding storing the mask. 
Currently, the constant mask is stored to memory and then immediately read back. Since the read address is often not aligned to the store address, this can cause an expensive Read-after-Write hazard. Although the load is contained entirely within the store, meaning that store forwarding can occur and there are no throughput penalties, there is still a latency penalty. The increased latency matters because of the pointer-chasing pattern in decompression here.

https://godbolt.org/z/e71jhE3bj

PiperOrigin-RevId: 626440765
 2026-05-09 18:52:39 +00:00
Danila Kutenin
c6d8c585dd Internal change 
PiperOrigin-RevId: 623073126
 2026-05-09 18:52:31 +00:00
Danila Kutenin
da1ea96391 Internal change 
PiperOrigin-RevId: 622802698
 2026-05-09 18:52:23 +00:00
Richard O'Grady
e1e7329dc7 Internal changes 
PiperOrigin-RevId: 621907514
 2026-05-09 18:52:12 +00:00
danilak-G
0212163f6e Merge pull request #234 from anthony-zy/rv_extractoffset 
RISC-V: Optimize decompression with branchless AdvanceToNextTagRVOptimized +13.7%
 2026-05-09 12:22:41 +01:00
danilak-G
0daaac5b07 Merge pull request #232 from anthony-zy/main 
refactor: Align RISC-V implementation with AArch64 branchless mask extraction
 2026-05-09 12:17:15 +01:00
danilak-G
70a0e1cbda Merge pull request #220 from yunfeizhou2025/upload 
limit RISC-V FindMatchLength optimizations to 64-bit
 2026-05-09 12:12:15 +01:00
anthony-zy
28590427a6 RISC-V: Optimize Snappy decompression tag advance (AdvanceToNextTagRVOptimized) for +4% throughput 2026-04-17 09:59:32 +08:00
郑紫阳10295048
77534271b3 refactor: Align RISC-V implementation with AArch64 branchless mask extraction 2026-04-15 10:33:35 +08:00
danilak-G
6281a07b7e Merge pull request #229 from UebelAndre/config 
Move `//:config` to `implementation_deps` to prevent header collision
 2026-03-06 22:30:42 +00:00
UebelAndre
4cb90917ee Move //:config to implementation_deps to prevent header collision 2026-03-06 11:47:10 -08:00
danilak-G
051dc4dc43 Merge pull request #228 from UebelAndre/bazel 
Add Bazel 9 support
 2026-03-03 16:49:56 +00:00
UebelAndre
07614b8308 Add Bazel 9 support 2026-02-28 12:38:26 -08:00
danilak-G
da459b5263 Create COPYING.notestdata 
This file is created for those who do not distribute testdata as it has different licenses, useful for some of our customers
 2026-01-21 10:12:49 +00:00
Zhou Yunfei
6484508ffb limit RISC-V FindMatchLength optimizations to 64-bit 
Add `__riscv_xlen == 64` check to ensure the optimized FindMatchLength
routine is only enabled on 64-bit RISC-V platforms.

Signed-off-by: Zhou Yunfei <xdzyf2012@gmail.com>
 2025-12-12 14:01:42 +08:00
danilak-G
25e52c58fb Merge pull request #214 from anthony-zy/add_zbb_check_macro 
RISC-V: gate __builtin_ctzll behind Zbb to avoid 10 % slowdown
 2025-11-20 12:13:53 +00:00
anthony-zy
a862a744b5 fix confic 2025-11-10 18:40:59 +08:00
anthony-zy
992fb9fcb0 Resolve merge conflicts 2025-11-10 18:39:54 +08:00
anthony-zy
c633c6bcb7 Resolve merge conflicts 2025-11-10 18:37:35 +08:00
anthony-zy
30dbdc2728 add zbb check macro 2025-11-10 18:31:13 +08:00
anthony-zy
727f5b3fc6 add zbb check macro 2025-11-10 10:37:19 +08:00
Danila Kutenin
cbea40d40c Fix an unused var warning for RISC-V 2025-10-20 13:34:53 +00:00
danilak-G
633cb9cbcb Merge pull request #213 from anthony-zy/add_rvv_optmized_memcopy64 
feat(RISC-V): Add RVV-optimized implementation for memcopy64
 2025-10-20 08:50:42 +01:00
anthony-zy
e92cb6ae8a Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-20 10:20:21 +08:00
anthony-zy
2d6c8f17de Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-20 10:19:39 +08:00
anthony-zy
b63569f403 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-20 09:53:37 +08:00
anthony-zy
9cc1596bf6 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-20 09:53:27 +08:00
anthony-zy
c098f689b8 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-17 17:53:04 +08:00
anthony-zy
6ab5882822 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-17 17:52:35 +08:00
anthony-zy
d457ac7380 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-17 10:29:45 +08:00
anthony-zy
9fd2b72c77 fix by some comments 2025-10-17 10:28:17 +08:00
anthony-zy
77a61a4927 Merge branch 'add_rvv_optmized_memcopy64' of https://github.com/anthony-zy/snappy into add_rvv_optmized_memcopy64 2025-10-17 10:21:26 +08:00
anthony-zy
c19d705ee1 Merge branch 'add_rvv_support' of https://github.com/anthony-zy/snappy into add_rvv_support 2025-10-17 10:20:57 +08:00
anthony-zy
2b8136579d Merge branch 'add_rvv_support' of https://github.com/anthony-zy/snappy into add_rvv_support 2025-09-12 11:07:09 +08:00
anthony-zy
c860cabbc5 Merge branch 'google:main' into add_rvv_support 2025-09-12 10:59:56 +08:00
anthony-zy
441032643f Merge branch 'google:main' into add_rvv_support 2025-09-05 16:02:38 +08:00
anthony-zy
27d8915d34 Merge branch 'add_rvv_support' of https://github.com/anthony-zy/snappy into add_rvv_support 2025-09-05 16:00:09 +08:00
anthony-zy
eb8d19fc1e Merge branch 'add_rvv_support' of https://github.com/anthony-zy/snappy into add_rvv_support 2025-09-05 15:59:29 +08:00
anthony-zy
0f989a2a97 Merge branch 'add_rvv_support' of https://github.com/anthony-zy/snappy into add_rvv_support 2025-08-28 20:42:49 +08:00
anthony-zy
802ef73aed add RVV support and optmized uncompress speed 2025-08-28 20:42:13 +08:00
anthony-zy
1f6ba69d67 add RVV support and optimized uncompress speed 2025-08-28 20:11:39 +08:00
anthony-zy
758804a4bc Use high-performance FindMatchLength to Optimize Snappy compression speed For RISC-V[skip ci] 2025-08-26 20:51:45 +08:00
danilak-G
6f99459b5b Merge pull request #208 from wanghan-sanechips/riscv-benchmark-fix 
build: Update benchmark submodule for RISC-V
 2025-07-29 09:45:06 +01:00
danilak-G
4793b4d004 Merge pull request #210 from anthony-zy/rv_findmatchlenght 
Use high-performance FindMatchLength to Optimize Snappy compression s…
 2025-07-29 09:44:35 +01:00
anthony-zy
1d88de7de8 Use high-performance FindMatchLength to Optimize Snappy compression speed For RISC-V[skip ci] 2025-07-29 14:08:43 +08:00
wanghan
b68435f083 build: Update benchmark submodule for RISC-V 
- Pins to b20cea67 to fix illegal instruction
- Verified on openEuler RISC-V
 2025-07-28 18:23:35 +08:00
Danila Kutenin
6af9287fbd Release 1.2.2 2025-03-26 15:19:22 +00:00
danilak-G
de19405e58 Merge pull request #203 from haney/msvc_shared 
Export symbols from snappy_test_support when building shared
 2025-03-26 15:18:20 +00:00
David Haney
4dd78f2deb Export symbols from snappy_test_support when building shared 
This replicates existing logic that was previously applied to the
`snappy` library.

fixes #202
 2025-03-26 08:10:25 -07:00